Bad weather and UK delivery availability
Restaurant operators blame the weather for soft trading more often than the weather deserves, and occasionally far less than it deserves. Both errors come from the same gap. No UK platform publishes what it does when a storm reduces courier supply, so whatever happens to your listing happens without explanation. From inside the building, a platform-side response to weather, a broken listing and a genuinely quiet night are indistinguishable, and each of the three calls for something different.
What do UK platforms say they do in bad weather?
Nothing we have been able to find, and that absence is worth stating rather than filling in.
What operators report observing falls into three shapes, escalating, each hitting a listing differently. Delivery estimates lengthen, which reduces conversion while leaving you visible. The area a customer can order from appears to contract, which makes you invisible on the edges while remaining normal to those nearby. Or ordering stops in an area altogether.
The middle one is the interesting case, because from the restaurant’s point of view nothing is wrong. The listing is live, orders are arriving, and a slice of the usual catchment simply cannot see you. Treat all three as patterns worth checking for rather than as published policy, because none of them is.
How would I know a zone had narrowed?
Not from the portal, in most cases, and not from your orders on the night.
Coverage is a property of what a customer at a given address sees, and there is no partner view that shows it as a map over time. A restaurant sees fewer orders from the outer postcodes, which is exactly what it would see if those customers had stayed in and cooked.
The way to know is to check availability from real points across the catchment and not from the listing alone. Kitchain (kitchain.co) measures coverage from points on the map, so a zone that shrank on the Tuesday and came back on the Thursday leaves a trace instead of a suspicion.
Is a weather pause the same as being closed?
Commercially, for the period it lasts, yes. Operationally it is different because you did not choose it and cannot lift it.
That matters for how you respond. There is no setting to correct and no support ticket that will restore a service reduced on the platform’s side. What is worth doing is knowing that it happened, so that the night is explained instead of investigated later.
It also matters for the numbers. Whether a platform-imposed pause counts against a restaurant’s availability metrics is not something any of them publishes, and it is a fair question to put to an account manager with dates attached.
How do I stop blaming weather for things that were not weather?
By having a comparison that weather cannot produce.
Weather is regional. A storm affects every site in a city, and mostly every brand. A listing failure affects one site. If three of your branches in one city were soft and one was catastrophic, the catastrophic one is not the weather.
That comparison needs per site data on the same night, which most groups can assemble from sales but not from availability. Availability is the version that answers the question cleanly, because it shows whether the site was orderable rather than whether anybody ordered.
Should hours change in bad weather?
Rarely, and less often than sites want.
A site that closes on the app because it expects a quiet night removes itself from the customers who were going to order anyway, and those are the highest intent customers of the week. Weather suppresses walk-in far more than delivery.
Where closing is genuinely right, because staff cannot get in or the kitchen cannot operate, do it as a deliberate closure with a set duration, not by switching off a device. A closure with an end time comes back on its own. A device left off does not.
What is the honest summary for a head office?
Weather is a real cause of lost orders and a poor explanation for a specific bad night at a specific site.
Treat it as a factor you can rule out rather than a conclusion you can reach. Ruling it out takes one comparison across sites and one look at whether the listing was actually orderable, and both are quick once somebody is keeping the record.